|
(11) | EP 2 620 944 B1 |
| (12) | EUROPEAN PATENT SPECIFICATION |
|
|
| (54) |
Asynchronous sampling frequency conversion device, method, and program Asynchrone Abtastfrequenzumwandlungsvorrichtung, Verfahren und Programm Dispositif de conversion de fréquence d'échantillonnage asynchrone, procédé et programme |
|
|
|||||||||||||||||||||||||||||||
| Note: Within nine months from the publication of the mention of the grant of the European patent, any person may give notice to the European Patent Office of opposition to the European patent granted. Notice of opposition shall be filed in a written reasoned statement. It shall not be deemed to have been filed until the opposition fee has been paid. (Art. 99(1) European Patent Convention). |
BACKGROUND
[0001] The present invention relates to an asynchronous sampling frequency conversion device, a method, and a program.
[0002] When an analog signal is converted into a digital signal, a sampling method of reading the waveform level of the analog signal at given time intervals and recording the read level value is used. In digital audio devices or the like, 32 kHz, 44.1 kHz, 48 kHz, and 96 kHz are generally used as sampling frequencies. For example, in cases of a CD (compact disc) and a DVD (digital versatile disc), sampling is performed at 44.1 kHz and 48 kHz, respectively.
[0003] For example, when a digital signal is transmitted and received between devices in which sampling frequencies are different, it is necessary to convert the sampling frequency of the digital signal. In order to change the sampling frequency of a digital signal, an asynchronous sampling rate converter is used.
[0004] A conventional asynchronous sampling rate converter changes a sampling frequency by performing up-sampling (0 compensation) on an input signal up to the least common multiple between sample frequencies before and after conversion, and then performing down-sampling (thinning) to a desired sampling frequency. Accordingly, for example, when the conventional asynchronous sampling rate converter converts a sampling frequency from 48 kHz to 44.1 kHz, the conventional asynchronous sampling rate converter has to perform up-sampling up to 7056 kHz which is the least common multiple between 48 kHz and 44.1 kHz, that is, up-sampling by 147 multiples of the input signal.
[0005] In the conventional asynchronous sampling rate converter, a LPF (LowPass Filter) that cuts a high-frequency component equal to or greater than the nyquist frequency (1/2 of the sampling frequency of an input signal) is disposed between an up-sampling unit and a down-sampling unit. An input signal into the LPF is a signal subjected to the up-sampling, and the sampling frequency of the signal is very high, as described above. For this reason, the LPF is required to have high-speed responsiveness and a high processing performance.
[0006] As a technology for suppressing a processing amount in an asynchronous sampling converter, a method of using linear interpolation is suggested by Japanese Laid-open Patent Publication No. 1-77326.
[0007] According to the method of using linear interpolation disclosed in Japanese Laid-open Patent Publication No. 1-77326, sample data Yi (where i is a natural number) after conversion is calculated from two pieces of sampling data Xn and Xn+1 (where n is a natural number) before conversion sampled at timings immediately before and after the sampling timing of Yi. Specifically, Yi is calculated by a calculation expression "Yi = (1 - ki)·Xn + ki·Xn + 1" using ki as an interpolation coefficient corresponding to Yi.
[0008] The interpolation coefficient ki is calculated by "ki = (i·θ2 - n·θ1)/θ1" by first specifying n that satisfies "n·θ1 ≤ i·θ2 < (n + 1)·θ1" for each i, when it is assumed that θ1 is an input sampling period and θ2 is an output sampling period. Thus, in the method of using the linear interpolation disclosed in Japanese Laid-open Patent Publication No. 1-77326, it is necessary to first specify "n" and calculate each interpolation coefficient ki using the specified "n" in order to calculate each sampling data Yi after conversion. For this reason, it takes much time to perform the calculation, since a calculation amount is large.
[0009] According to the sampling theorem, an input signal with a frequency band up to 1/2 of an input clock frequency fin(= 1/θ1) can be reproduced. However, due to the frequency characteristics of the linear interpolation, the amplitude of a waveform reproduced from the output sampling data Yi calculated through the linear interpolation decreases, as the frequency band of an input signal increases. Therefore, waveform distortion may be worsened. For example, with regard to an input signal with a frequency band of 1/2 of an input clock frequency fin, the amplitude of a waveform formed by the output sampling data Yi may decrease about 0.64 times the amplitude of the input signal.
[0010] Attention is drawn to JP H06 204 798 A, which shows an interpolation system for an asynchronous sampling frequency conversion. In the system, the data interval of input data a1 based on an original sampling frequency is divided equally by a prescribed number M, a filtered value is interpolated at an equally division point just before or just after an output phase point with respect to an asynchronous sampling frequency and closest to the output phase point among equally division points and Lagrangian interpolation is executed by using three data composed of the result of the filter interpolation, input data a-1 just before the output phase point, and input data al just after the output phase point. The result of the Lagrangian interpolation is used for output data (b) at the output phase point.
SUMMARY
[0011] In accordance with the present invention, an asynchronous sampling frequency conversion device as set forth in claim 1, an asynchronous sampling frequency conversion method as set forth in claim 5 and an asynchronous sampling frequency conversion program as set forth in claim 6 are provided. Further embodiments of the invention are inter alia disclosed in the dependent claims.
[0012] According to an aspect of the present invention, an asynchronous sampling frequency conversion device inter alia includes: a storage unit (104) configured to store input digital signals; a data specifying unit (105A) configured to specify first data and second data based on a ratio of a sampling frequency of the input digital signal to a sampling frequency of an output digital signal, the first data being sampled at a sampling timing immediately before an ith (where i is a natural number) sampling timing of the output digital signal among the input digital signals stored in the storage unit (104), the second data being sampled at the sampling timing immediately after the ith sampling timing of the output digital signal; and an output data value calculator (105C) configured to calculate a value of ith data of the output digital signal based on the first data and the second data specified by the data
specifying unit (105A) and the ratio.
BRIEF DESCRIPTION OF THE DRAWINGS
[0013]
FIG. 1 is a block diagram illustrating an example of an asynchronous sampling frequency conversion device according to a first embodiment;
FIG. 2 is a diagram illustrating an example of input sampling data information of an L channel according to the first embodiment;
FIG. 3 is a flowchart illustrating a process of generating the input sampling data information of the L channel according to the first embodiment;
FIG. 4 is a flowchart illustrating an L channel conversion process according to the first embodiment;
FIG. 5 is a diagram illustrating a correspondence relation among a sequence i of output sampling, a product (i·fin/fOUT) of the sequence i and a ratio fin/fOUT, a rounddown (i·fin/fOUT), a roundup (i·fin/fOUT), input sampling data Xn, and input sampling data Xn+1, when a process of converting a sine wave with 8 kHz is performed according to the first embodiment;
FIG. 6 is a diagram illustrating a list of the output sampling data of the L channel according to the first embodiment;
FIG. 7 is a diagram illustrating a waveform reproduced from the output sampling data illustrated in FIG. 6 and a waveform reproduced from input sampling data of the L channel;
FIG. 8 is a block diagram illustrating an example of an asynchronous sampling frequency conversion device according to a second embodiment;
FIG. 9 is a diagram illustrating a list of the output sampling data of an L channel according to the second embodiment;
FIG. 10 is a diagram illustrating a waveform reproduced from the output sampling data illustrated in FIG. 9 and a waveform reproduced from input sampling data of the L channel; and
FIG. 11 is a diagram illustrating an example of output sampling information of the L channel.
DETAILED DESCRIPTION
[0014] An asynchronous sampling frequency conversion device according to preferred embodiments of the invention will be described.
First Embodiment
[0015] An asynchronous sampling frequency conversion device 1 according to a first embodiment is a device that converts the sampling frequency of an input digital signal and outputs the converted sampling frequency. As illustrated in FIG. 1, the asynchronous sampling frequency conversion device 1 includes an input signal receiver 101, an LPF 102, a storage processing unit 103, a storage (a storage unit) 104, a sampling frequency conversion processing unit 105, an output signal transmitter 106, and clock generating circuits 107 and 108.
[0016] For example, the input signal receiver 101 receives (latches) an input digital signal Sin corresponding to two channels of an L channel and an R channel sent with I2S (The IC Sound Bus) which is a communication standard of a digital sound signal, left justified 24 bits, or the like in response to an RL clock RLin supplied from the clock generating circuit 107, and outputs the input digital signal Sin to the LPF 102. For example, the RL clock RLin is generated from an RL clock transmitted from a signal supply source and has the same frequency as that of a sampling frequency (hereinafter, referred to as an input sampling frequency) fin of the input digital signal Sin. When the input sampling frequency fin is, for example, 48 kHz, the input signal receiver 101 receives two pieces of input sampling data of the L channel and the R channel for each about 20.8 µsec (= 1/48 kHz).
[0017] The LPF (LowPass Filter) unit 102 operates in response to the RL clock RLin supplied from the clock generating circuit 107 and limits the bandwidth of the input digital signal Sin. More specifically, the LPF 102 removes waveform distortion caused due to a return phenomenon by cutting a high-frequency wave component equal to or greater than the nyquist frequency (1/2 of input sampling frequency fin = fin/2) of the input signal Sin.
[0018] The storage processing unit 103 includes a CPU and a memory that stores an operation program of the CPU. The storage processing unit 103 includes an internal counter 103C with an initial setting value of "0." The CPU included in the storage processing unit 103 executes the operation program, counts a number (sequence) n of data (input sampling data) X output from the LPF 102 by counting up the internal counter 103C in response to the RL clock signal RLin, and calculates a time Tn from sampling start to an nth sampling timing, that is, n/fin. The storage processing unit 103 forms input data by causing the sequence n, the elapsed time Tn (= n/fin), and the nth input sampling data Xn to correspond to each other, and stores a predetermined number of input data in the storage 104. When the predetermined number of input data are stored in the storage 104, the storage processing unit 103 erases the oldest (where the sequence n is the newest) input data and stores new input data. Further, data stored in the storage 104 is referred below to as input sampling data information.
[0019] For example, when the input digital signal Sin is assumed to be data obtained by sampling an analog sine wave sin (2π·8000·t) with 8 kHz using the sampling frequency fin with 48 kHz, the storage processing unit 103 sequentially stores the sequence n, the elapsed time Tn (= n/fin), and the input sampling data Xn in the storage 104, as illustrated in FIG. 2.
[0020] The storage 104 includes a RAM (Random Access Memory) and stores up to the predetermined number of input data in which the sequence n, the elapsed time Tn (= n/fin), and the nth input sampling data Xn generated by the storage processing unit 103 correspond to each other. The storage 104 has a capacity in which a predetermined number of input data necessary for the sampling frequency conversion process performed by the sampling frequency conversion processing unit 105 can be stored. When the input sampling frequency is 48 kHz, for example, the storage 104 stores 480 pieces of input data of each of the L channel and the R channel. The number of input data retained in the storage 104 is merely an example. For example, the storage 104 may be configured to store 960 pieces of input data which is a double of the 480 pieces of input data.
[0021] The sampling frequency conversion processing unit 105 includes a CPU and an internal memory. The internal memory functions as a work area of a CPU (Central Processing Unit) and stores a conversion processing program executed by the CPU and fixed data necessary for the conversion process. The CPU functions as a data specifying unit 105A, a determining unit 105B, an output data value calculator 105C, and a counter 105D by executing the conversion processing program and calculates output sampling data of an output sampling frequency fout using the input sampling data information stored in the storage 104 by performing a conversion process to be described below.
[0022] The data specifying unit 105A specifies input sampling data sampled at timings immediately before and after the sampling timing of output sampling data Yi (where i is an output sequence and is a natural number). Here, "the timings immediately before and after the output sampling timing" does not mean physical timings, but means sampling timings of the input digital signal Sin immediately before and after the output sampling timing by specifying a timing of the input digital signal Sin corresponding to the output sampling timing in consideration of delay of signal processing and using the specified timing as a reference.
[0023] Specifically, the data specifying unit 105A acquires a ratio fin/fout of the input sampling frequency fin to the output sampling frequency fout. For example, when the data specifying unit 105A calculates the ith output sampling data Yi of the output digital signal, the data specifying unit 105A calculates a value roundup (i·fin/fout) obtained by rounding up a value (i·fin/fout), which is i times the acquired ratio fin/fout, to a natural number and a value rounddown (i·fin/fout) obtained by rounding down a value (i·fin/fout), which is i times the acquired ratio fin/fout, to a natural number. For example, when i·fin/fout = 1.12, the roundup (i·fin/fout) = 2 and the rounddown (i·fin/fout) = 1.
[0024] When the rounddown (i·fin/fout) = n (natural number), the data specifying unit 105A specifies an nth input sampling data Xn as the input data sampled at timings immediately before the sampling timing of the output sampling data Yi, and specifies an (n + 1)th input sampling data Xn+1 as the input data sampled at timings immediately after the sampling timing of the output sampling data Yi.
[0025] Here, when the rounddown (i·fin/fout) = the roundup (i·fin/fout) = n, that is, a value (i·fin/fout), which is i times the ratio fin/fout, is a natural number, the data specifying unit 105A specifies only the input sampling data Xn sampled at the same timing as the sampling timing of the output sampling data Yi as the data sampled at the timings immediately before and after the sampling timing of the output sampling data Yi.
[0026] The determining unit 105B determines whether the rounddown (i·fin/fout) = the roundup (i·fin/fout) = n. That is, the determining unit 105B determines whether a value (i·fin/fout) which is i times the ratio fin/fout is a natural number. Further, the determining unit 105B determines whether there are input sampling data corresponding to the rounddown (i·fin/fout) and the roundup (i·fin/fout) calculated by the data specifying unit 105A. Specifically, the determining unit 105B determines whether there are the sampling sequences corresponding to the rounddown (i·fin/fout) and the roundup (i·fin/fout) by searching for a column of the sampling sequence of the input sampling data information stored in the storage 104.
[0027] The output data value calculator 105C calculates the output sampling data Yi using the input sampling data Xn and Xn+1 specified by the data specifying unit 105A. Specifically, the output data value calculator 105C weights the input sampling data Xn with respect to a value obtained by subtracting the (i·fin/fout) from the roundup (i·fin/fout), and weights the input sampling data Xn+1 with respect to a value obtained by subtracting the rounddown (i·fin/fout) from (i·fin/fout) . Then, the output data value calculator 105C calculates the sum of the weighted input sampling data as the ith output sampling data Yi of the output digital signal. That is, the output data value calculator 105C calculates the output sampling data Yi by the following expression (1).
[0028] In this expression, {(i·fin/fout) - rounddown (i·fin/fout)} corresponds to an interpolation coefficient ki in linear interpolation and {roundup (i·fin/fout) - (i·fin/fout)} corresponds to (1 - ki).
[0029] Here, the determining unit 105B determines that the product (i·fin/fout) is a natural number, that is, the rounddown (i·fin/fout) = the roundup (i·fin/fout) = n, the output data value calculator 105C calculates the input sampling data Xn specified by the data specifying unit 105A as the output sampling data Yi. Then, the sampling frequency conversion processing unit 105 outputs the output sampling data Yi calculated by the output data value calculator 105C to the output signal transmitter 106.
[0030] The counter 105D increments a counter value, whenever the output data value calculator 105C calculates the output sampling data Yi. The initial setting value of the counter value of the counter 105D is "1." The counter value of the counter 105D indicates a sampling sequence i in the output sampling frequency. The data specifying unit 105A specifies the input sampling data Xn and Xn+1 based on the product (i·fin/fout) of the counter value i of the counter 105D and the ratio fin/fout.
[0031] The output signal transmitter 106 includes an FIFO memory. The output signal transmitter 106 stores the output sampling data Yi supplied from the sampling frequency conversion processing unit 105 and sequentially outputs the output sampling data Yi as the output digital signal Sout in synchronization with the RL clock RLout of the output digital signal Sout supplied from the clock generating circuit 108.
[0032] The clock generating circuit 107 supplies the RL clock RLin synchronized with the RL clock RLin supplied from a source supplying the input digital signal Sin to the input signal receiver 101, the LPF 102, and the storage processing unit 103.
[0033] The clock generating circuit 108 generates and outputs the RL clock RLout of the output digital signal Sout. The RL clock RLout has the same frequency as the sampling frequency fout of the output digital signal Sout.
[0034] Next, a sampling frequency conversion process of the asynchronous sampling frequency conversion device 1 having the above-described configuration will be described.
[0035] Further, in the sampling frequency conversion process, a process on the data of the L channel and a process on the data of the R channel process are performed in parallel. However, the process for the L channel will be described below, since the process for the L channel is the same as the process of the R channel.
[0036] The sampling frequency conversion process includes a process of receiving the input digital signal Sin and storing the input sampling data information in the storage 104 and a process of generating the output digital signal Sout using the input sampling data information stored in the storage 104.
[0037] First, the process of receiving the input digital signal Sin and then storing the input sampling data information in the storage processing unit 103 will be described with reference to FIG. 3.
[0038] The input sampling data (signal pieces) forming the input digital signal Sin are sequentially supplied to the input signal receiver 101. The input signal receiver 101 sequentially receives (latches) the supplied input sampling data (signal pieces) in response to the RL clock RLin and sequentially supplies the input sampling data to the LPF 102.
[0039] The LPF 102 operates in response to the RL clock RLin, filters the input sampling data supplied from the input signal receiver 101 by removing a high-frequency wave component from the the input sampling data, and supplies the result to the storage processing unit 103.
[0040] The storage processing unit 103 starts a process of generating the input sampling data information illustrated in FIG. 3 in response to the RL clock RLin supplied from the clock generating circuit 107.
[0041] First, the storage processing unit 103 (accurately, the CPU of the storage processing unit 103) increments the counter value (sequence) n of the internal counter 103C (step S01) and calculates the sequence n of the input sampling data.
[0042] Next, the storage processing unit 103 calculates the elapsed time Tn (= n/fin) from the sampling start to an nth sampling timing by calculating a product of the counter value (sequence) n and a supply period 1/fin of the input digital signal Sin (step S02).
[0043] Next, the storage processing unit 103 acquires the input sampling data Xn output at the nth time from the LPF 102 (step S03) and determines whether the number of acquired input sampling data Xn is greater than a predetermined number (step S04).
[0044] When the storage processing unit 103 determines that the number of acquired input sampling data Xn is greater than the predetermined number (YES in step S04), the storage processing unit 103 wastes the oldest stored input data among the input sampling data information stored in the storage 104 from the input sampling data information (step S05). Then, the storage processing unit 103 adds input data formed by the counter value (sequence) n of the internal counter 103C incremented in the process of step S01, the elapsed time Tn calculated in the process of step S02, and the input sampling data Xn acquired in the process of step S03 to the input sampling data information (step S06) .
[0045] Conversely, when the storage processing unit 103 determines that the number of acquired input sampling data Xn is not greater than the predetermined number (NO in step S04), the process of step S05 is skipped, and then the storage processing unit 103 adds input data formed by the counter value (sequence) n of the internal counter 103C incremented in the process of step S01, the elapsed time Tn calculated in the process of step S02, and the input sampling data Xn acquired in the process of step S03 to the input sampling data information (step S06).
[0046] By repeating the above-described process of generating the input sampling data information, as illustrated in FIG. 2, a predetermined number of input sampling data X immediately adjacent to each other are stored as the input sampling data information in the storage 104.
[0047] Next, a conversion process performed by the sampling frequency conversion processing unit 105 will be described with reference to FIG. 4. The sampling frequency conversion processing unit 105 performs conversion process for the L channel and the R channel in parallel. The conversion process for the L channel will be described below, since the conversion process for the L channel is the same as the conversion process for the R channel. The conversion process starts using RLout as a trigger.
[0048] It is supposed that the ratio fin/fout of the input sampling frequency fin to the output sampling frequency fout is calculated in advance and is stored in a non-volatile memory or the like included in the internal memory of the sampling frequency conversion processing unit 105.
[0049] The data specifying unit 105A acquires the ratio fin/fout of the input sampling frequency fin to the output sampling frequency fout stored in the internal memory of the sampling frequency conversion processing unit 105 (step 001). Then, the data specifying unit 105A acquires the counter value i from the counter 105D (step 002), and then calculates the product (i·fin/fout) of the ratio fin/fout acquired in step 001 and the counter value i (step 003). Then, the data specifying unit 105A calculates the rounddown (i·fin/fout) obtained by rounding down the value (i·fin/fout) calculated in step 003 and the roundup (i·fin/fout) obtained by rounding up the value (i·fin/fout) (step 004).
[0050] Then, the determining unit 105B determines whether the value n of the rounddown (i·fin/fout) is the same as the value n of the roundup (i·fin/fout) by comparing the rounddown (i·fin/fout) to the roundup (i·fin/fout). That is, the determining unit 105B determines whether the product (i·fin/fout) calculated in step 003 is a natural number (step 005). When the determining unit 105B determines that the product (i·fin/fout) is a natural number (YES in step 005), the determining unit 105B searches for the column of the sampling sequence of the input sampling data information stored in the storage 104 and determines whether there is the nth input sampling data Xn corresponding to the value of n of the rounddown (i·fin/fout) calculated in step 004 (step 021).
[0051] When the determining unit 105B determines that there is the nth input sampling data Xn corresponding to the value of n of the rounddown (i·fin/fout) calculated in step 004 (YES in step 021), the data specifying unit 105A specifies this input sampling data Xn as the data sampled at the timings immediately before and after the sampling timing of the output sampling data Yi (step 022). Then, the output data value calculator 105C calculates the input sampling data Xn specified by the data specifying unit 105A as the output sampling data Yi and the sampling frequency conversion processing unit 105 outputs the output sampling data Yi (input sampling data Xn) calculated by the output data value calculator 105C to the output signal transmitter 106 (step 023). Then, the process proceeds to step 009.
[0052] Conversely, when the determining unit 105B determines that there is no nth input sampling data Xn (NO in step 021), the sampling frequency conversion processing unit 105 initializes the counter value of the counter 105D (step 041) and ends the conversion process.
[0053] Conversely, when the determining unit 105B determines that the product (i·fin/fout) is not a natural number (NO in step 005), the determining unit 105B searches for the column of the sampling sequence of the input sampling data information stored in the storage 104 and determines whether there are both the nth input sampling data Xn corresponding to the value n of the rounddown (i·fin/fout) calculated in step 004 and the (n + 1)th input sampling data Xn+1 (step 006). When the determining unit 105B determines that there is at least no input sampling data Xn+1 (NO in step 006), the process proceeds to step 041. Then, the sampling frequency conversion processing unit 105 initializes the counter value of the counter 105D, deletes the input sampling data information stored in the storage 104 (step 041), and ends the conversion process.
[0054] Conversely, when the determining unit 105B determines there are both the input sampling data Xn and the input sampling data Xn+1 (YES in step 006), the data specifying unit 105A specifies the input sampling data Xn and Xn+1 (step 007). Then, the output data value calculator 105C calculates the output sampling data Yi by the calculation expression (Expression (1)) of the above-described output sampling data Yi using the input sampling data Xn and Xn+1 specified by the data specifying unit 105A and the sampling frequency conversion processing unit 105 outputs the output sampling data Yi calculated by the output data value calculator 105C to the output signal transmitter 106 (step 008). Then, the sampling frequency conversion processing unit 105 increments the counter value of the counter 105D (step 009) and returns to step 001 to repeat the above-described processes.
[0055] Referring to FIGS. 2, 4, and 5, the process of each unit will be described using 48 kHz and 44.1 kHz as the input sampling frequency fin and the output sampling frequency fout and giving examples of specific numerical values. As in FIG. 2, FIG. 5 is a diagram illustrating a correspondence relation among the sampling sequence i at an output sampling frequency, the product (i·fin/fout) of the sequence i and the ratio fin/fout, the rounddown (i·fin/fout), the roundup (i·fin/fout), and the input sampling data Xn and Xn+1 in the case in which the rounddown (i·fin/fout) and the roundup (i·fin/fout) are equal to n, when the conversion process according to this embodiment is performed on a sine wave sin (2π·8000·t) with 8 kHz.
[0056] The data specifying unit 105A acquires the ratio fin/fout stored in the internal memory of the sampling frequency conversion processing unit 105 and the counter value i of the counter 105D (steps 001 and 002). Then, the data specifying unit 105A calculates a product of the ratio fin/fout and the counter value i (step 003).
[0057] For example, when the counter value i is "1" and i·fin/fout = 1 × (48 ÷ 44.1) = 1.088435374, the data specifying unit 105A calculates 1.088435374 as the value of i·fin/fout.
[0058] Next, the data specifying unit 105A calculates the rounddown (i·fin/fout) and the roundup (i·fin/fout) (step 004). When i·fin/fout = 1.088435374, the data specifying unit 105A calculates each of "the rounddown (i·fin/fout) = 1" and "the roundup (i·fin/fout) = 2."
[0059] Next, the determining unit 105B determines whether the product (i·fin/fout) is a natural number (step 005). When it is assumed that i·fin/fout = 1.088435374, NO is determined in step 005. Then, the determining unit 105B determines whether there are both the first (rounddown (i·fin/fout) = 1) input sampling data X1 and the second (roundup (i·fin/fout) = 2) input sampling data X2 in the input sampling data information stored in the storage 104 (step 006). When it is supposed that the input sampling data information stored in the storage 104 is information illustrated in FIG. 2, the determining unit 105B determines that there are both the input sampling data X1 and X2 (YES in step 006) and the data specifying unit 105A specifies "X1 = 0.5" and "X2 = 0.866025404" (step 007).
[0060] Next, the output data value calculator 105C applies the data values of the input sampling data X1 and X2 specified by the data specifying unit 105A, the value of the product (i·fin/fout), and the values of the rounddown (i·fin/fout) and the roundup (i·fin/fout) to the above-described calculation expression (Expression (1)), calculates the value of the ith output sampling data Yi of the output digital signal, and outputs the calculated value to the output signal transmitter 106 (step 008). In the example, since i = 1, the first output sampling data Y1 calculated by the output data value calculator 105C is "{0.5 × (2 - 1.088435374) + 0.866025404 × (1.088435374 - 1)} = 0.53237." Next, the sampling frequency conversion processing unit 105 increments the counter value of the counter 105D (step 009), returns to the process of step 001, and sequentially calculates the output sampling data Yi by repeating the series of processes.
[0061] Thus, the sampling frequency conversion processing unit 105 sequentially calculates the output sampling data Yi for the L channel and the R channel at the output sampling frequency using the calculation expression (Expression (1)) of the above-described output sampling data Yi, and sequentially supplies the output sampling data Yi to the output signal transmitter 106.
[0062] Further, the sampling frequency conversion processing unit 105 preferably generates the output sampling data Yi at a processing speed at which one piece of data is calculated at one period of the RL clock RLout or a speed slightly faster than the processing speed.
[0063] The output signal transmitter 106 stores the output sampling data Yi supplied from the sampling frequency conversion processing unit 105 in the FIFO memory and sequentially outputs the output sampling data Yi in synchronization with the RL clock RLout of the output signal supplied from the clock generating circuit 108.
[0064] As described above, by configuring the asynchronous sampling frequency conversion device 1 according to this embodiment, it is possible to easily specify the input sampling data Xn and Xn+1 sampled immediately before and after the sampling timing of the output sampling data Yi based on the product (i·fin/fout) of the counter value i of the counter 105D and the ratio fin/fout stored in the internal memory of the sampling frequency conversion processing unit 105. Further, as understood from the calculation expression (Expression (1)) of the output sampling data Yi according to this embodiment, the output sampling data Yi can be calculated by calculating the product (i·fin/fout) of the counter value i of the counter 105D and the ratio fin/fout, rounding down and rounding up the fraction of the calculated product to calculate natural numbers n and n + 1, specifying the input sampling data Xn and Xn+1 based on the calculated natural numbers n and n + 1, and using the data values. Accordingly, the calculation amount necessary to calculate the output sampling data Yi can be reduced compared to the conventional linear interpolation method of specifying "n" indicating the sampling sequence of the input sampling data Xn sampled immediately before the sampling timing of the output sampling data Yi and calculating each interpolation coefficient ki using the "n." Thus, the conversion process in the asynchronous sampling frequency conversion device 1 can be performed at high speed.
[0065] The LPF 102 may operate at the input sampling frequency fin of the input digital signal Sin and may operate at a relatively low frequency.
[0066] The input sampling data Xn is once stored in the storage 104. Therefore, even when jitter is included in data to be input, the jitter can be absorbed.
Second Embodiment
[0067] In the first embodiment, the output data value calculator 105C of the sampling frequency conversion processing unit 105 has calculated the output sampling data Yi by the above-described calculation expression (Expression (1)). FIG. 6 is a diagram illustrating output sampling data Yi of an L channel at the output sampling frequency of 44.1 kHz calculated by the output data value calculator 105C according to the first embodiment. FIG. 7 is a diagram illustrating a waveform Wout reproduced from the output sampling data Yi illustrated in FIG. 6, that is, the output sampling data Yi of the L channel at the output sampling frequency of 44.1 kHz calculated by the output data value calculator 105C and a waveform Win reproduced from the input sampling data Xn of the L channel at the input sampling frequency of 48 kHz according to the first embodiment.
[0068] When the waveform Wout and the waveform Win are compared to each other with reference to FIG. 7, the amplitude of the waveform Wout is reduced more than that of the waveform Win. The reduction in the amplitude of the waveform Wout is caused due to the frequency characteristics of the linear interpolation, as described in Background. Accordingly, in this embodiment, a waveform close to the waveform Win is configured to be reproduced from the output sampling data by using the output sampling data Yi calculated in the first embodiment as a reference value and correcting the reference value Yi using a change amount |Xn - Xn+1| between the input sampling data Xn and Xn+1. Thus, it is possible to remove a waveform distortion caused due to a decrease in the amplitude originated from the frequency characteristics of the linear interpolation.
[0069] As illustrated in FIG. 8, an asynchronous sampling frequency conversion device 1' according to this embodiment is basically the same configuration as the asynchronous sampling frequency conversion device of the first embodiment, since the asynchronous sampling frequency conversion device 1' includes an input signal receiver 101, an LPF 102, a storage processing unit 103, a storage 104, a sampling frequency conversion processing unit 105, and an output signal transmitter 106. However, the sampling frequency conversion processing unit 105 further includes a change amount calculator 105E and a correcting unit 105F. Further, the internal memory of the sampling frequency conversion processing unit 105 stores a correction coefficient α included in a calculation expression for output sampling data Zi to be described below according to this embodiment. The correction coefficient α is a value experimentally calculated in advance. For example, the correction coefficient α is determined by calculating a decrease ratio of the amplitude caused due to the frequency characteristics of the linear interpolation in a frequency band of an input signal. Further, the correction coefficient α is preferably determined so that a waveform clip does not occur.
[0070] The change amount calculator 105E calculates a change amount |Xn - Xn+1| between the input sampling data Xn and Xn+1 specified by the data specifying unit 105A.
[0071] The correcting unit 105F uses the output sampling data Yi calculated by the output data value calculator 105C as a reference value and corrects the reference value Yi based on the change amount |Xn - Xn+1| calculated by the change amount calculator 105E. More specifically, the correcting unit 105F corrects the reference value Yi by the following calculation expression and calculates the corrected value as the output sampling data Zi.
[0072] The sampling frequency conversion processing unit 105 outputs the value Zi corrected by the correcting unit 105F as an ith output sampling data to the output signal transmitter 106.
[0073] FIG. 9 is a diagram illustrating a list of the output sampling data Zi calculated by performing the conversion process on a sine wave sin (2π·8000·t) with 8 kHz at the input sampling frequency of 48 kHz and the output sampling frequency of 44.1 kHz, when the correction coefficient α is set to 0.2. FIG. 10 is a diagram illustrating a waveform WZout reproduced from the output sampling data Zi illustrated in FIG. 9, that is, the output sampling data Zi of the L channel at the output sampling frequency of 44.1 kHz calculated by the correcting unit 105F and a waveform Win reproduced from the input sampling data Xn of the L channel at the input sampling frequency of 48 kHz.
[0074] When the waveform WZout is compared to the waveform Win with reference to FIG. 10, it can be understood that the waveform WZout is closer to the waveform Win, compared to FIG. 7 illustrating the waveform Wout and the waveform Win reproduced from the output sampling data Yi calculated in the first embodiment. This can be also understood from the output sampling data Yi and Zi indicated in FIGS. 6 and 9, respectively. For example, when Y3 in FIG. 6 is compared to Z3 in FIG. 9, the decrease in the amplitude originated from the frequency characteristics of the linear interpolation is improved so that the amplitude of Z3 is closer to the amplitude "1" of the sine wave sin (2π·8000·t) with 8 kHz since Y3 = 0.964456 and Z3 = 0.990298.
[0075] According to the method of calculating the output sampling data Zi using the change amount |Xn - Xn+1| between the input sampling data Xn and Xn+1, as described in this embodiment, the decrease ratio of the amplitude originated from the frequency characteristics of the linear interpolation increases, as the frequency of an input signal increases. Therefore, since the waveform distortion is accordingly noticeable, a high effect can be obtained.
[0076] In the above-described first and second embodiments, the cases have been described in which the input sampling frequency fin is 48 kHz and the output sampling frequency fout is 44.1 kHz, but the invention is not limited thereto. The input sampling frequency fin and the output sampling frequency fout are arbitrarily set including the magnitude relation therebetween. For example, the input sampling frequency fin may be set to 48 kHz and the output sampling frequency fout may be set to 32 kHz. Further, the input sampling frequency fin may be set to 44.1 kHz or 32 kHz and the output sampling frequency fout may be set to 48 kHz.
[0077] In the conversion process performed by the sampling frequency conversion processing unit 105, when the sampling frequency is lowered, the LPF 102 is necessary in order to set the input digital signal Sin to be equal to or less than the nyquist frequency. However, when the sampling frequency is raised, the LPF 102 is not necessary.
[0078] The above-described first and second embodiments of the invention can be also applied to an asynchronous sampling frequency conversion device in which the input sampling frequency fin is varied. In this case, the sampling frequency conversion processing unit 105 may be configured to calculate the ratio fin/fout of the input sampling frequency fin to the output sampling frequency fout. Since the sampling frequency generally used in a digital device is limited, a ratio of a sampling frequency to the output sampling frequency fout may be retained in advance in the internal memory of the sampling frequency conversion processing unit 105.
[0079] In the above-described first and second embodiments, the cases have been described in which the internal memory of the sampling frequency conversion processing unit 105 stores the ratio fin/fout of the input sampling frequency fin to the output sampling frequency fout in advance, but the invention is not limited thereto. The sampling frequency conversion processing unit 105 may calculate the ratio fin/fout.
[0080] The example has been described in which the storage processing unit 103 calculates the elapsed time Tn and stores the elapsed time Tn in the storage 104. However, the sampling frequency conversion processing unit 105 may calculates the elapsed time Tn.
[0081] In the above-described first embodiment, the storage 104 may be configured to beforehand retain output sampling information formed by a pair of "i" indicating the sampling sequence of the output sampling data Yi, a product (i·fin/fout) of the ratio fin/fout of the input sampling frequency fin to the output sampling frequency fout and "i" indicating the sampling sequence of the output sampling data Yi, and the value n of the rounddown (i·fin/fout), as exemplified in FIG. 11. Thus, since the data specifying unit 105A can specify the value n of the rounddown (i·fin/fout) from the counter value i of the counter 105D indicating the sampling sequence of the output sampling data Yi, the calculation time of the output sampling data can be further shortened. Therefore, the conversion process performed by the sampling frequency conversion processing unit 105 can be performed at higher speed.
[0082] In the above-described second embodiment, the case has been described in which the calculation expression (Expression (2)) of the output sampling data Zi is Zi = Yi·(1 + α·|Xn - Xn+1|), but the invention is not limited thereto. The output sampling data Zi may be calculated by the calculation expression of "Zi = Yi·{(1 + β·F(|Xn - Xn+1|)}" by using F(x) and β as a function and a correction coefficient, respectively.
[0083] In the above-described embodiments, the above-described processes may be performed by storing and distributing a program (a computer program product) to be executed in a computer-readable recording medium such as a flexible disk, a CD-ROM (Compact Disc Read-Only Memory), DVD, or an MO (Magneto-Optical Disc) and installing the program.
[0084] Further, a program may be stored in a disc device or the like included in a predetermined server apparatus on a communication network such as the Internet and may be acquired, for example, by superimposing the program on carrier waves.
[0085] In the above-described embodiments, it is possible to reduce a calculation amount and also perform a conversion process at high speed. Furthermore, it is possible to realize excellent reproducibility of a signal waveform.
[0086] Although the invention has been described with respect to specific embodiments for a complete and clear disclosure, the appended claims are not to be thus limited but are to be construed as embodying all modifications and alternative constructions that may occur to one skilled in the art that fairly fall within the basic teaching herein set forth.
a storage unit (104) configured to store input digital signals;
a data specifying unit (105A) configured to specify first data and second data based on a ratio of a sampling frequency of the input digital signal to a sampling frequency of an output digital signal, the first data being sampled at a sampling timing immediately before an ith, where i is a natural number, sampling timing of the output digital signal among the input digital signals stored in the storage unit (104), the second data being sampled at the sampling timing immediately after the ith sampling timing of the output digital signal; and
an output data value calculator (105C) configured to calculate a value of ith data of the output digital signal based on the first data and the second data specified by the data specifying unit (105A) and the ratio, characterized in that the data specifying unit (105A) is configured to specify the first data that is rounddown value in accord with a sequence of the sampling timing, and to specify the second data that is roundup value in accord with the sequence of the sampling timing, the rounddown value being obtained by rounding down a multiplication value to a natural number, the roundup value being obtained by rounding up the multiplication value to a natural number, the multiplication value being obtained by multiplying the ratio and i indicating the sequence of the sampling timing of data to be calculated by the output data value calculator (105C); and
the output data value calculator (105C) is configured to calculate the value of the ith data of the output digital signal through linear interpolation using the first data and the second data specified by the data specifying unit (105A), the multiplication value, the rounddown value, and the roundup value.
a storage processing unit (103) configured to store the input digital signals in correspondence with sequences of the sampling timing and values of sequentially sampled data in the storage unit (104),
wherein the data specifying unit (105A) is configured to search for the sequence of the sampling timing which corresponds to the sampled data stored in the storage unit (104), to specify, as the first data, data corresponding to the sequence of the sampling timing in accord with the rounddown value and to specify, as the second data, data corresponding to the sequence of the sampling timing in accord with the roundup value.
a determining unit (105B) configured to determines whether the multiplication value is a natural number,
wherein, when the determining unit (105B) determines that the multiplication value is the natural number,
the data specifying unit (105A) is configured to specify, as the first data and the second data, data sampled at the same timing as the ith sampling timing of the output digital signal among the input digital signals stored in the storage unit (104), and
the output data value calculator (105C) is configured to calculate, as the value of the ith data of the output digital signal, a value of data sampled at the same timing as the ith sampling timing of the output digital signal specified by the data specifying unit (105A).
a change amount calculator (105E) configured to calculate an absolute value of a change amount of the second data with respect to the first data; and
a correcting unit (105F) configured to correct the value of the ith data of the output digital signal calculated by the output data value calculator (105C) based on the absolute value of the change amount of the second data with respect to the first data, which is calculated by the change amount calculator (105E).
a signal storing step for storing input digital signals;
a data specifying step for specifying first data and second data based on a ratio of a sampling frequency of the input digital signal to a sampling frequency of an output digital signal, the first data being sampled at a sampling timing immediately before an ith, where i is a natural number, sampling timing of the output digital signal among the input digital signals stored in the signal storing step, the second data being sampled at the sampling timing immediately after the ith sampling timing of the output digital signal; and
a data calculating step for calculating a value of ith data of the output digital signal based on the specified first data, the specified second data and the ratio,
characterized in that,
in the data specifying step, the first data that is rounddown value in accord with a sequence of the sampling timing and the second data that is roundup value in accord with the sequence of the sampling timing are specified, the rounddown value being obtained by rounding down a multiplication value to a natural number, the roundup value being obtained by rounding up the multiplication value to a natural number, the multiplication value being obtained by multiplying the ratio and i indicating the sequence of the sampling timing of data to be calculated in the data calculating step; and
in the data calculating step, the value of the ith data of the output digital signal is calculated through linear interpolation using the first data and the second data specified in the data specifying step, the multiplication value, the rounddown value, and the roundup value.
a signal process for storing input digital signals;
a data specifying step for specifying first data and second data based on a ratio of a sampling frequency of the input digital signal to a sampling frequency of an output digital signal, the first data being sampled at a sampling timing immediately before an ith, where i is a natural number, sampling timing of the output digital signal among the input digital signals stored in the signal storing process, the second data being sampled at the sampling timing immediately after the ith sampling timing of the output digital signal; and
a data calculating step for calculating a value of ith data of the output digital signal based on the specified first data, the specified second data and the ratio, characterized in that, in the data specifying process, the first data that is rounddown value in accord with a sequence of the sampling timing and the second data that is roundup value in accord with the sequence of the sampling timing are specified, the rounddown value being obtained by rounding down a multiplication value to a natural number, the roundup value being obtained by rounding up the multiplication value to a natural number, the multiplication value being obtained by multiplying the ratio and i indicating the sequence of the sampling timing of data to be calculated in the data calculating process; and
in the data calculating process, the value of the ith data of the output digital signal is calculated through linear interpolation using the first data and the second data specified in the data specifying process, the multiplication value, the rounddown value, and the roundup value.
eine Speichereinheit (104), die konfiguriert ist, um Eingangsdigitalsignale zu speichern;
eine Datenspezifizierungseinheit (105A), die konfiguriert ist, um erste Daten und zweite Daten basierend auf einem Verhältnis einer Abtast- bzw. Samplingfrequenz des Eingangsdigitalsignals zu einer Samplingfrequenz des Ausgangsdigitalsignals zu spezifizieren, wobei die ersten Daten zu einem Abtast- bzw. Samplingzeitpunkt direkt von einem i-ten Samplingzeitpunkt des Ausgangsdigitalsignals gesampled bzw. abgetastet werden, wobei i eine natürliche Zahl ist, und zwar unter den Eingangsdigitalsignalen, die in der Speichereinheit (104) gespeichert sind, wobei die zweiten Daten zu dem Samplingzeitpunkt direkt nach dem i-ten Samplingzeitpunkt des Ausgangsdigitalsignals gesampled bzw. abgetastet werden; und
einen Ausgangsdatenwertkalkulator (105C), der konfiguriert ist, um einen Wert der der i-ten Daten des Ausgangsdigitalsignals basierend auf den ersten Daten und den zweiten Daten, die von der Datenspezifizierungseinheit (105A) spezifiziert sind, und dem Verhältnis zu berechnen,
dadurch gekennzeichnet, dass
die Datenspezifizierungseinheit (105A) konfiguriert ist, um die ersten Daten zu spezifizieren, die ein Abrundungswert sind, und zwar in Übereinstimmung mit einer Abfolge des Samplingzeitpunktes, und um die zweiten Daten zu spezifizieren, die ein Aufrundungswert sind, und zwar in Übereinstimmung mit der Abfolge des Samplingzeitpunktes, wobei der Abrundungswert erhalten wird durch Abrunden eines Multiplikationswertes auf eine natürliche Zahl, wobei der Aufrundungswert erhalten wird durch Aufrunden des Multiplikationswertes auf eine natürliche Zahl, wobei der Multiplikationswert erhalten wird durch Multiplizieren des Verhältnisses mit i, was die Abfolge des Samplingzeitpunktes der Daten anzeigt, die von dem Ausgangdatenwertkalkulator (105C) berechnet werden sollen; und
wobei der Ausgangdatenwertkalkulator (105C) konfiguriert ist, um den Wert der i-ten Daten des Ausgangsdigitalsignals durch lineare Interpolation unter Verwendung der ersten Daten und der zweiten Daten, die durch die Datenspezifizierungseinheit (105A) spezifiziert wurden, des Multiplikationswertes, des Abrundungswertes und des Aufrundungswertes zu berechnen.
eine Speicherverarbeitungseiheit (103), die konfiguriert ist, um die Eingangsdigitalsignale in Übereinstimmung mit Sequenzen der Samplingzeit und Werten von sequentiell gesampelten bzw. abgetasteten Daten in der Speichereinheit (104) zu speichern,
wobei die Datenspezifizierungseinheit (105A) konfiguriert ist, um nach der Sequenz des Samplingzeitpunktes zu suchen, die den gesampelten Daten entspricht, die in der Speichereinheit (104) gespeichert sind, um als die ersten Daten Daten zu spezifizieren, welche der Sequenz des Samplingzeitpunktes entsprechen, und zwar in Übereinstimmung mit dem Abrundungswert, und um als die zweiten Daten Daten zu spezifizieren, die der Sequenz des Samplingzeitpunktes entsprechen, und zwar in Übereinstimmung mit dem Aufrundungswert.
eine Bestimmungseinheit (105B), die konfiguriert ist, um zu bestimmen, ob der Multiplikationswert eine natürliche Zahl ist,
wobei, wenn die Bestimmungseinheit (105B) bestimmt, dass der Multiplikationswert die natürliche Zahl ist,
die Datenspezifizierungseinheit (105A) konfiguriert ist, um als die ersten Daten und die zweiten Daten Daten zu spezifizieren, die zu dem gleichen Zeitpunkt gesampelt bzw. abgetastet wurden, wie zu dem i-ten Samplingzeitpunkt des Ausgangsdigitalsignals, und zwar unter den Eingangsdigitalsignalen, die in der Speichereinheit (104) gespeichert sind, und
wobei der Ausgangdatenwertkalkulator (105C) konfiguriert ist, um als den Wert der i-ten Daten des Ausgangsdigitalsignals einen Wert von Daten zu berechnen, die zum gleichen Zeitpunkt gesampelt wurden wie zu dem iten Samplingzeitpunkt des Ausgangsdigitalsignals, welches durch die Datenspezifizierungseinheit (105A) spezifiziert wurde.
einen Veränderungsgrößenkalkulatur (105E), der konfiguriert ist, um einen absoluten Wert einer Veränderungsgröße der zweiten Daten bezüglich der ersten Daten zu berechnen; und
eine Korrektureinheit (105F), die konfiguriert ist, um den Wert der i-ten Daten des Ausgangsdigitalsignals, welches von dem Ausgangdatenwertkalkulator (105C) berechnet wurde, zu korrigieren, und zwar basierend auf dem absoluten Wert der Veränderungsgröße der zweiten Daten bezüglich der ersten Daten, der berechnet wird durch den Veränderungsgrößenkalkulatur (105E).
einen Signalspeicherschritt zum Speichern von Eingangsdigitalsignalen;
einen Datenspezifizierungsschritt zum Spezifizieren von ersten Daten und zweiten Daten basierend auf einem Verhältnis einer Abtast- bzw. Samplingfrequenz des Eingangsdigitalsignals zu einer Samplingfrequenz eines Ausgangsdigitalsignals, wobei die ersten Daten zu einem Samplingzeitpunkt direkt vor einem i-ten Samplingzeitpunkt des Ausgangsdigitalsignals gesampelt bzw. abgetastet werden, wobei i eine natürliche Zahl ist, und zwar unter den Eingangsdigitalsignalen, die in dem Signalspeicherschritt gespeichert werden, wobei die zweiten Daten zu dem Samplingzeitpunkt direkt nach dem i-ten Samplingzeitpunkt des Ausgangsdigitalsignals gesampelt werden; und
einen Datenberechnungsschritt zum Berechnen eines Wertes von i-ten Daten des Ausgangsdigitalsignals basierend auf den spezifizierten ersten Daten, den spezifizierten zweiten Daten und dem Verhältnis,
dadurch gekennzeichnet, dass
in dem Datenspezifizierungsschritt die ersten Daten, die ein Abrundungswert in Übereinstimmung mit einer Sequenz des Samplingzeitpunktes sind, und die zweiten Daten, die ein Aufrundungswert in Übereinstimmung mit der Sequenz des Samplingzeitpunktes sind, spezifiziert werden, wobei der Abrundungswert erhalten wird durch Abrunden eines Multiplikationswertes auf eine natürliche Zahl, wobei der Aufrundungswert erhalten wird durch Aufrunden des Multiplikationswertes auf eine natürliche Zahl, wobei der Multiplikationswert erhalten wird durch Multiplizieren des Verhältnisses mit i, was die Sequenz des Samplingzeitpunktes der Daten anzeigt, die in dem Datenberechnungsschritt berechnet werden sollen; und
wobei in dem Datenberechnungsschritt der Wert der i-ten Daten ds Ausgangsdigitalsignals berechnet wird durch Linearinterpolation unter Verwendung der ersten Daten und der zweiten Daten, die in dem Datenspezifizierungsschritt spezifiziert wurden, des Multiplikationswertes, des Abrundungswertes und des Aufrundungswertes.
einen Signalprozess zum Speichern von Eingangsdigitalsignalen; einen Datenspezifizierungsschritt zum Spezifizieren von ersten Daten und zweiten Daten basierend auf einem Verhältnis einer Abtast- bzw. Samplingfrequenz des Eingangsdigitalsignals zu einer Samplingfrequenz eines Ausgangsdigitalsignals, wobei die ersten Daten zu einem Samplingzeitpunkt direkt vor einem i-ten Samplingzeitpunkt des Ausgangsdigitalsignals abgetastet bzw. gesampelt werden, wobei i eine natürliche Zahl ist, und zwar unter den Eingangsdigitalsignalen, die in dem Signalspeicherprozess gespeichert wurden, wobei die zweiten Daten zu einem Samplingzeitpunkt direkt nach dem i-ten Samplingzeitpunkt des Ausgangsdigitalsignals gesampelt werden; und
einen Datenberechnungsschritt zum Berechnen eines Wertes der i-ten Daten des Ausgangsdigitalsignals basierend auf den spezifizierten ersten Daten, den spezifizierten zweiten Daten und dem Verhältnis,
dadurch gekennzeichnet, dass
in dem Datenspezifizierungsprozess die ersten Daten, die ein Abrundungswert in Übereinstimmung mit einer Sequenz des Samplingzeitpunktes sind, und die zweiten Daten, die ein Aufrundungswert in Übereinstimmung mit der Sequenz des Samplingzeitpunktes sind, spezifiziert werden, wobei der Abrundungswert erhalten wird durch Abrunden eines Multiplikationswertes auf eine natürliche Zahl, wobei der Aufrundungswert erhalten wird durch Aufrunden des Multiplikationswertes auf eine natürliche Zahl, wobei der Multiplikationswert erhalten wird durch Multiplizieren des Verhältnisses mit i, was die Sequenz des Samplingzeitpunktes von Daten anzeigt, die in dem Datenberechnungsprozess berechnet werden sollen; und in dem Datenberechnungsprozess der Wert der i-ten Daten des Ausgangsdigitalsignals durch Linearinterpolation berechnet wird, und zwar unter Verwendung der ersten Daten und der zweiten Daten, die in dem Datenspezifizierungsprozess spezifiziert wurden, des Multiplikationswertes, des Abrundungswertes und des Aufrundungswertes.
un module de stockage (104) agencé pour mémoriser des signaux numériques d'entrée ;
un module de spécification de données (105A) agencé pour spécifier des premières données et des deuxièmes données sur la base d'un rapport entre une fréquence d'échantillonnage du signal numérique d'entrée et une fréquence d'échantillonnage d'un signal numérique de sortie, les premières données étant échantillonnées à un instant d'échantillonnage situé immédiatement avant un instant d'échantillonnage i, i étant un nombre naturel, du signal numérique de sortie parmi les signaux numériques d'entrée mémorisés dans le module de stockage (104), les deuxièmes données étant échantillonnées à l'instant d'échantillonnage situé immédiatement après l'instant d'échantillonnage i du signal numérique de sortie ; et
un calculateur de valeur de données de sortie (105C) agencé pour calculer une valeur de données i du signal numérique de sortie sur la base des premières données et des deuxièmes données spécifiées par le module de spécification de données (105A) et du rapport,
caractérisé en ce que
le module de spécification de données (105A) est agencé pour spécifier les premières données qui sont une valeur arrondie vers le bas en accord avec une séquence d'instants d'échantillonnage, et pour spécifier les deuxièmes données qui sont une valeur arrondie vers le haut en accord avec la séquence d'instants d'échantillonnage, la valeur arrondie vers le bas étant obtenue en arrondissant vers le bas une valeur de multiplication pour obtenir un nombre naturel, la valeur arrondie vers le haut étant obtenue en arrondissant vers le haut la valeur de multiplication pour obtenir un nombre naturel, la valeur de multiplication étantobtenue en multipliant le rapport et i indiquant la séquence d'instants d'échantillonnage de données à calculer par le calculateur de valeur de données de sortie (105C) ; et
le calculateur de valeur de données de sortie (105C) est agencé pour calculer la valeur de données i du signal numérique de sortie par une interpolation linéaire en utilisant les premières données et les deuxièmes données spécifiées par le module de spécification de données , la valeur de multiplication, la valeur arrondie vers le bas, et la valeur arrondie vers le haut.
un module de traitement de stockage (103) agencé pour mémoriser les signaux numériques d'entrée en correspondance avec des séquences d'instants d'échantillonnage et des valeurs de données échantillonnées séquentiellement dans le module de stockage (104),
dans lequel le module de spécification de données (105A) est agencé pour rechercher la séquence d'instants d'échantillonnage qui correspond aux données échantillonnées mémorisées dans le module de stockage (104), pour spécifier, comme première données, des données correspondant à la séquence d'instants d'échantillonnage en accord avec la valeur arrondie vers le bas et pour spécifier, comme deuxièmes données, des données correspondant à la séquence d'instants d'échantillonnage en accord avec la valeur arrondie vers le haut.
unmodule de détermination (105B) agencé pour déterminer si la valeur de multiplication est un nombre naturel,
dans lequel, lorsque le module de détermination (105B) détermine que la valeur de multiplication est un nombre naturel,
le module de spécification de données (105A) est agencé pour spécifier, comme premières données et comme deuxièmes données, des données échantillonnées au même instant que l'instant d'échantillonnage i du signal numérique de sortie parmi les signaux numériques d'entrée mémorisés dans le module de stockage (104), et
le calculateur de valeur de données de sortie (105C) est agencé pour calculer, comme valeur de données i du signal numérique de sortie, une valeur de données échantillonnée au même instant que l'instant d'échantillonnage i du signal numérique de sortie spécifié par le module de spécification de données (105A).
un calculateur de quantité de changement (105E) agencé pour calculer une valeur absolue d'une quantité de changement des deuxièmes données par rapport aux premières données ; et
un module de correction (105F) agencé pour corriger la valeur de données i du signal numérique de sortie calculé par le calculateur de valeur de données de sortie (105C) sur la base de la valeur absolue de la quantité de changement des deuxièmes données par rapport aux premières données, qui sont calculées par le calculateur de quantité de changement (105E).
une étape de mémorisation de signal pour mémoriser des signaux numériques d'entrée ;
une étape de spécification de données pour spécifier des premières données et des deuxièmes données sur la base d'un rapport entre une fréquence d'échantillonnage du signal numérique d'entrée et une fréquence d'échantillonnage d'un signal numérique de sortie, les premières données étant échantillonnées à un instant d'échantillonnage situé immédiatement avant l'instant d'échantillonnage i, i étant un nombre naturel, du signal numérique de sortie parmi les signaux numériques d'entrée mémorisés dans l'étape de mémorisation de signal, les deuxièmes données étant échantillonnées à l'instant d'échantillonnage situé immédiatement après l'instant d'échantillonnage i du signal numérique de sortie ; et
une étape de calcul de données pour calculer une valeur des données i du signal numérique de sortie sur la base des premières données spécifiées, des deuxièmes données spécifiées et du rapport,
caractérisé en ce que,
dans l'étape de spécification de données, les premières données qui sont une valeur arrondie vers le bas en accord avec une séquence d'instants d'échantillonnage et les deuxièmes données qui sont une valeur arrondie vers le haut en accord avec la séquence d'instants d'échantillonnage sont spécifiées, la valeur arrondie vers le bas étant obtenue en arrondissant vers le bas une valeur de multiplication pour obtenir un nombre naturel, la valeur arrondie vers le haut étant obtenue en arrondissant vers le haut la valeur de multiplication pour obtenir un nombre naturel, la valeur de multiplication étant obtenue en multipliant le rapport et i indiquant la séquence d'instants d'échantillonnage de données à calculer dans l'étape de calcul de données ; et
dans l'étape de calcul de données, la valeur de données i du signal numérique de sortie est calculée par une interpolation linéaire utilisant les premières données et les deuxièmes données spécifiées dans l'étape de spécification de données, la valeur de multiplication, la valeur d'arrondi vers le bas et la valeur d'arrondi vers le haut.
un traitement de signal pour mémoriser des signaux numériques d'entrée ;
une étape de spécification de données pour spécifier des premières données et des deuxièmes données sur la base d'un rapport entre une fréquence d'échantillonnage du signal numérique d'entrée et une fréquence d'échantillonnage d'un signal numérique de sortie, les premières données étant échantillonnées à un instant d'échantillonnage situé immédiatement avant l'instant d'échantillonnage i, i étant un nombre naturel, du signal numérique de sortie parmi les signaux numériques d'entrée mémorisés dans l'étape de mémorisation de signal, les deuxièmes données étant échantillonnées à l'instant d'échantillonnage situé immédiatement après l'instant d'échantillonnage i du signal numérique de sortie ; et
une étape de calcul de données pour calculer une valeur des données i du signal numérique de sortie sur la base des premières données spécifiées, des deuxièmes données spécifiées et du rapport,
caractérisé en ce que,
dans le processus de spécification de données, les premières données qui sont une valeur arrondie vers le bas en accord avec une séquence d'instants d'échantillonnage et les deuxièmes données qui sont une valeur arrondie vers le haut en accord avec la séquence d'instants d'échantillonnage sont spécifiées, la valeur arrondie vers le bas étant obtenue en arrondissant vers le bas une valeur de multiplication pour obtenir un nombre naturel, la valeur arrondie vers le haut étant obtenue en arrondissant vers le haut la valeur de multiplication pour obtenir un nombre naturel, la valeur de multiplication étant obtenue en multipliant le rapport et i indiquant la séquence d'instants d'échantillonnage de données à calculer dans le processus de calcul de données ; et
dans le processus de calcul de données, la valeur de données i du signal numérique de sortie est calculée par une interpolation linéaire utilisant les premières données et les deuxièmes données spécifiées dans le processus de spécification de données, la valeur de multiplication, la valeur d'arrondi vers le bas et la valeur d'arrondi vers le haut.
REFERENCES CITED IN THE DESCRIPTION
This list of references cited by the applicant is for the reader's convenience only. It does not form part of the European patent document. Even though great care has been taken in compiling the references, errors or omissions cannot be excluded and the EPO disclaims all liability in this regard.
Patent documents cited in the description